Venice AI, the privacy-first AI platform run by crypto entrepreneur Erik Voorhees, has closed its first outside funding round: $65 million at a $1 billion post-money valuation. The July 1 deal was led by Dragonfly, the crypto-native venture firm, with Coinbase Ventures, North Island Ventures, Archetype, Liquid2 Ventures, Morgan Creek, and F-Prime Capital also participating.

The figures that matter most aren't in the cap table. Venice entered this raise already profitable, with annualized revenue passing $70 million, 3.5 million registered users, and roughly 2 million daily API calls across more than 200 AI models. None of that revenue came from training on user conversations, because Venice doesn't store them.

Two Years Without a Server Log

Venice launched in May 2024 without a dollar of outside investment. Voorhees, who previously founded Satoshi Dice and the ShapeShift cryptocurrency exchange, built the platform around a single architectural decision: user prompts are encrypted client-side and routed through an external proxy before hitting any model, and Venice keeps no record of the interaction on its own systems.

Erik Voorhees, Venice AI CEO, says the company turned profitable "while choosing not to" track user conversations. Source: decrypt.co

The business model is straightforward: paid subscribers get access to the full model catalog and end-to-end encryption. The platform also offers a crypto payment layer through its dual-token system - VVV and DIEM - where staking VVV creates daily AI credits. Only 8% of users actually pay in crypto, Voorhees has acknowledged, which means the token architecture is more about philosophy than transaction volume.

The company handles conversations across more than 200 models, a mix of open-source options like Llama and Mistral alongside routed access to closed-source providers. It processed 1.3 trillion tokens last month at around 2 million daily API calls. That scale puts it well ahead of most privacy-focused AI experiments, which tend to stay small precisely because privacy constraints add friction.

Who Is Actually Buying This

On the metrics that traditional venture prefers, Venice looks like a discount. A 14x revenue multiple is well below the 20-30x that OpenAI and Anthropic command despite neither being profitable. The investor thesis here isn't that Venice competes with those platforms head-to-head. It's that a sizeable population of users cares about where their prompts go, and that population is willing to pay.

The investor lineup reinforces that bet. Dragonfly, Coinbase Ventures, and the other participants are not generalist tech funds stumbling into AI. They're crypto-native operators who built their theses on user sovereignty and censorship resistance. Venice is that same thesis applied to a different infrastructure layer: instead of who controls your money, who controls your thought process.

"Intelligence is becoming a collaboration between man and machine. Venice's mission is to protect it from mass surveillance and censorship," Voorhees said in the funding announcement.

The company plans to use the capital to purchase GPUs outright and build its own data centers. That move tightens the privacy argument: instead of leasing compute where a cloud provider could theoretically examine workloads, Venice would control the hardware running the queries. It's also a margin play - owned infrastructure is cheaper over multi-year horizons than renting from AWS or CoreWeave.

This approach mirrors what similar AI funding deals have done - vertical integration as both a cost and a defensibility strategy.

Venice AI's interface routes queries through client-side encryption before they reach any model. Source: venice.ai

Who Benefits

Privacy-conscious users get the most obvious upside. Venice's architecture means the platform can't train on your questions about medical symptoms, legal problems, or anything else you'd rather keep off a server log. For the subset of AI users who have read the terms of service on ChatGPT or Claude - and don't love what they found - Venice offers a concrete alternative rather than a marketing claim. Contrast this with cases like XChat's encryption promises, where the keys still lived on the company's servers.

The open-source AI ecosystem also benefits indirectly. Venice routes traffic to open models as well as closed ones, and its growth validates the consumer market for non-OpenAI model access. Every user who builds a habit on Venice is a user who isn't locked into a single provider's model family.

Developers using the API get a compliant, privacy-preserving inference endpoint that doesn't require them to trust a major lab with their users' data. For companies in healthcare, legal, and financial services - where data handling is regulated - that's a meaningful product distinction rather than a nice-to-have. This contrasts with OpenAI's approach, which added a privacy filter as an optional layer on top of a system that still collects data by default.

Who Pays

The three major labs - OpenAI, Anthropic, and Google - lose the segment of users who make a deliberate choice against surveillance. That cohort may not be large as a percentage of the total AI market, but it skews toward power users and developers who tend to spend more and switch less.

Cloud GPU providers face a different risk. Venice's plan to build owned compute is small in absolute terms now, but it signals a broader trend: AI companies that start on rented infrastructure eventually buy hardware when they can afford to. Every dollar Venice spends at a GPU fab is a dollar that doesn't go to CoreWeave or AWS.

VVV token holders have a mixed picture. The buy-and-burn program - Venice uses subscription revenue to purchase and destroy VVV tokens - provides support. But the real value of the token depends on Venice maintaining platform growth. If the company hits a ceiling on privacy-motivated users, the token economics get harder to sustain.

The $65 million round values Venice at 14x revenue. That multiple implies the investors believe the market for private AI is larger than it looks now, or that Venice can expand beyond its current user base without abandoning the privacy constraints that define its product. Those two things are hard to do at the same time.

Sources:

• Venice AI Blog - Series A Announcement
• TechCrunch - Venice AI becomes a unicorn with $65M Series A
• Decrypt - Venice AI Valued at $1 Billion
• CryptoBriefing - Venice AI hits $1B valuation as Erik Voorhees bets big on privacy
• SiliconAngle - Venice raises $65M at $1B valuation